Cost-effectiveness of embryo transfer strategies: a decision analytic model using long-term costs and consequences of singletons and multiples born as a consequence of IVF

What is the cost-effectiveness of elective single embryo transfer (eSET) versus double embryo transfer (DET) strategies from a societal perspective, when applying a time horizon of 1, 5 and 18 years? From a short-term perspective (1 year) it is cost-effective to replace DET with single embryo transfer; however when intermediate- (5 years) and long-term (18 years) costs and consequences are incorporated, DET becomes the most cost-effective strategy, given a ceiling ratio of a,not sign20 000 per quality-adjusted life years (QALY) gained. According to previous cost-effectiveness research into embryo transfer strategies, DET is considered cost-effective if society is willing to pay around a,not sign20 000 for an extra live birth. However, interpretation of those studies is complicated, as those studies fail to incorporate long-term costs and outcomes and used live birth as a measure of effectiveness instead of QALYs. With this outcome, both multiple and singletons were valued as one live birth, whereas costs of all children of a multiple were incorporated. A Markov model (cycle length: 1 year; time horizon: 1, 5 and 18 years) was developed comparing a maximum of: (i) three cycles of eSET in all patients; (ii) four cycles of eSET in all patients; (iii) five cycles of eSET in all patients; (iv) three cycles of standard treatment policy (STP), i.e. eSET in women <38 years with a good quality embryo, and DET in all other women; and (v) three cycles of DET in all patients. Expected life years (LYs), child QALYs and costs were estimated for all comparators. Input parameters were derived from a retrospective cohort study, in which hospital resource data were collected (n=580) and a parental questionnaire was sent out (431 respondents). Probabilistic sensitivity analysis (5000 iterations) was performed. With a time horizon of 18 years, DETx3 is most effective (0.54 live births, 10.2 LYs and 9.8 QALYs) and expensive (a,not sign37 871) per couple starting IVF. Three cycles of eSET are least effective (0.43 live births, 7.1 LYs and 6.8 QALYs) and expensive (a,not sign25 563). We assumed that society is willing to pay a,not sign20 000 per QALY gained. With a time horizon of 1 year, eSETx3 was the most cost-effective embryo transfer strategy with a probability of being cost-effective of 99.9%. With a time horizon of 5 or 18 years, DETx3 was most cost-effective, with probabilities of being cost-effective of 77.3 and 93.2%, respectively. This is the first study to use QALYs generated by the children in the economic evaluation of embryo transfer strategies. There remains some disagreement on whether QALYs generated by new life should be used in economic evaluations of fertility treatment. A further limitation is that treatment ends when it results in live birth and that only child QALYs were considered as measure of effectiveness. The results for the time horizon of 18 years might be less solid, as the data beyond the age of 8 years are based on extrapolation. The current Markov model indicates that when child QALYs are used as measure of outcome it is not cost-effective on the long term to replace DET with single embryo transfer strategies. However, for a balanced approach, a family-planning perspective would be preferable, including additional treatment cycles for couples who wish to have another child. Furthermore, the analysis should be extended to include QALYs of family members. This study was supported by a research grant (grant number 80-82310-98-09094) from the Netherlands Organization for Health Research and Development (ZonMw). There are no conflicts of interest in connection with this article. Not applicable.